Author: Eelco de Groot
Date: 07:39:26 09/26/03
Go up one level in this thread
On September 26, 2003 at 05:01:22, Thorsten Czub wrote:
>the main problem is that is have to use
>a style based on 9375 for the slow programs (hiarcs, shredder)
>and on the base of 9344 for the fast programs.
>
>so far i was not able to create a style that works for both opponent
>programs :-((
It's a pity Thorsten. But the result against Shredder still stands!
I agree with Ed that [Pruning ? MISC_37] * New extension might not be so bad. I
remember that taking it out for some testpositions made results -for those
positions- worse. At least I believe in this extension more than in [Pruning =
MISC_40] * New reduction , which I suspect of making it harder to find the Bg7
defence that Thomas Lagershausen mentioned in
http://f50.parsimony.net/forum200336/messages/515.htm
At least that was the case for my settings. Possible improvements would probably
be a lot of work for Ed... For instance, if the history depth does best at 9,
would it do even better if you could set it at 10? And if so might it not be
possible that taking the history depth idea back and not remembering it in the
hashtable could be an idea?
Also if some extensions work best at blitz or best at the earlier plydepths, how
about for a new Rebel some extensions like [Pruning ? MISC_37] that could be
switched off after a certain plydepth? Maybe that would work for some reductions
too, like the Fail High Reductions but then to start them only at plydepth 12
for instance. But that is I think for the future...
Sorry Ed, maybe that would be an awful lot of work..
I would like it to be noted in the records that today I managed to beat Macheïde
Lite 0 just once, with very experimental settings! The only thing that the
settings have going for them -it's the only thing I have tested them with- is
that they play 43.e4-e5 in the Ferret - Gandalf game.
[D]rr4k1/5p2/3p2p1/3P3p/pnP1P2P/Bn2NR2/5PP1/qB1Q2K1 w - - 1 1
Rebel Betazoid Experimental Version 5so
00:00:00.4 -1,50 1 7 Bxb4 Rxb4
00:00:00.5 -0,95 1 134 Nc2
00:00:00.5 -2,55 2 217 Nc2 Nxc2 Bxd6
00:00:00.6 -1,13 2 345 Bxb4 Rxb4 Nc2
00:00:00.7 -0,24 3 2237 Bxb4 Nd4 Qe1 Nxf3+
00:00:00.9 -0,24 4 3438 Bxb4 Nd4 Qe1 Nxf3+ gxf3
00:00:01.0 -0,50 5 10506 Bxb4 Nd4 Qe1 Nxf3+ gxf3 Qb2
00:00:01.2 -1,08 6 27147 Bxb4 Nd4 Qe1 Qb2 Nc2 Nxc2
00:00:01.6 -1,42 7 79395 Bxb4 Nd4 Qe1 Qb2 Nc2 Nxc2 Bxc2 Rxb4
00:00:02.2 -1,42 8 191103 Bxb4 Nd4 Qe1 Qb2 Nc2 Nxc2 Bxc2 Rxb4
00:00:04.4 -1,99 9 626999 Bxb4 Nd4 Qe1 Qb2 Bxd6 Nxf3+ gxf3 Qxb1
00:00:09.7 -1,99 10 1728001 Bxb4 Nd4 Qe1 Qb2 Bxd6 Nxf3+ gxf3 Qxb1
00:00:22.1 -2,31 11 4526944 Bxb4 Nd4 Qe1 Qb2 Bxd6 Nxf3+ gxf3 Qxb1
00:01:03.7 -2,31 12 13578044 Bxb4 Nd4 Qe1 Qb2 Bxd6 Nxf3+ gxf3 Qxb1
00:04:24.5 -3,14 13 58749479 Bxb4 Nd4 Qe1 Qb2 Bxd6 Qxb1 Qxb1
00:10:47.9 -3,00 13 140894601 e5 Qxa3 e6 Qc1 Rxf7
This called for a testgame, if only because it was the first time I could get
Rebel to reject Bxb4 at ply 13...
It is not easy to beat the Macheïde!
[Event "*"]
[Site "Groningen"]
[Date "2003.09.26"]
[Round "1"]
[White "Rebel 12 pre-release - Betazoid 5so"]
[Black "Rebel 12 pre-release - Macheide Lite 0"]
[Result "1-0"]
[TimeControl "40/1200:40/1200"]
{ERT 0.9.0.4 Autoflag=On Pondering=Off OperatorTime=300 mS } 1. d4 {0,00/1 0}
Nf6
{0,00/1 0} 2. c4 {0,00/1 0} g6 {0,00/1 0} 3. Nc3 {0,00/1 0} Bg7 {0,00/1 0} 4. e4
{0,00/1 0} d6
{0,00/1 0} 5. f3 {0,00/1 0} O-O {0,00/1 0} 6. Bg5 {0,00/1 0} Nc6 {0,00/1 0} 7.
Qd2 {0,00/1 0} a6
{0,00/1 0} 8. Nge2 {0,00/1 0} Rb8 {0,00/1 0} 9. d5 {0,00/1 0} Ne5 {0,00/1 0} 10.
Nd4 {0,00/1 0}
c5 {0,00/1 0} 11. Nc2 {0,00/1 0} Nh5 {0,00/1 0} 12. Be2 {0,00/1 0} f5 {0,00/1 0}
13. exf5
{0,00/1 0} gxf5 {0,00/1 0} 14. f4 {0,00/1 0} Ng4 {0,59/11 47 O-O b5 Qe1 Qb6 cxb5
e5}
15. O-O {0,71/10 56 h6 Bh4 b5 Rae1 b4 Nd1 Nhf6} Bd7 {0,50/10 79 Rae1 Re8 h3} 16.
Rae1
{0,78/10 52 Nhf6 a4 b5 cxb5 axb5 Nxb5 Ne4} Re8 {0,63/10 50 h3} 17. h3
{0,91/10 59 Ngf6 Bh4 Bh6 a4 Kh8 b3} Ngf6 {0,76/11 74 Bh4} 18. Bh4 {0,88/10 48
Bh6 g3}
Bh8 {0,82/10 65 Qd1 Ng7 Bd3 b5} 19. Qd1 {1,02/10 64 Ng7 Bxf6 exf6 Bh5 Nxh5 Qxh5
b5}
Ng7 {0,68/11 61 Bd3 b5} 20. Bxf6 {0,96/10 51 exf6 Bd3 b5 Ne3 Qb6 g3} exf6
{0,65/9 7 Bd3}
21. Bd3 {0,91/10 39 b5 Ne3 b4 Na4 Re7 Bc2 Qe8} b5 {0,69/11 59 Rxe8+ Bxe8} 22.
Rxe8+
{0,99/10 57 Qxe8 cxb5 axb5 Re1 Qf7 a3} Bxe8 {0,69/10 15 Ne3 bxc4} 23. Ne3
{0,99/11 61 Qd7 Re1} Bh5 {0,69/11 57 Qb1 bxc4 Nxc4 Bg6 Re1 Qc7} 24. Qb1
{1,17/11 42 bxc4} bxc4 {0,91/10 27 Nxc4 Qc7 Bxf5 Nxf5 Qxf5 Bg6} 25. Nxc4
{1,21/11 30 Bg6 Re1 Qc7 g3 Kf8 Kh2} Qd7 {0,93/11 58 Re1 Re8 Rxe8+ Bxe8} 26. Ne3
{1,27/10 33 Qa7 Re1 Bg6 Nc4 Qd7 Qc2} Qa7 {1,06/11 65 Rf2 Kf8 Nxf5} 27. Rf2
{1,35/10 29 Qe7 Nxf5} c4 {1,09/11 46 Nxc4} 28. Nxc4
{1,70/11 53 Qc5 Nd2 Qb6 Bxf5 Nxf5 Qxf5 Bg6} Qc5 {1,40/11 58 Nd2 Qb6 Bxf5} 29.
Nd2
{1,80/10 27 Qb6 Bxf5 Nxf5 Qxf5 Bg6 Qe6+ Bf7} Qb6 {1,37/10 24 Bxf5} 30. Bxf5
{1,88/10 26 Nxf5 Qxf5 Bg6 Qe6+ Bf7 Qg4+} Nxf5 {1,59/11 30 Qxf5 Bg6} 31. Qxf5
{1,76/9 6 Bg6 Qe6+ Bf7 Qg4+} Bg6 {1,69/11 44 Qe6+ Bf7 Qg4+ Kf8} 32. Qe6+
{2,10/11 40 Bf7 Qg4+} Bf7 {1,69/10 29 Qg4+ Bg7 Nde4 Kf8 Kh2 Re8 Qd7} 33. Qg4+
{2,11/11 87 Bg7 Nde4 Qa7} Bg7 {1,76/10 26 Nde4 Kf8 Kh2 Re8 Re2 Qc7 f5} 34. Nde4
{2,07/10 38 Kf8 Kh2 Re8} Qa7 {1,83/10 29 Kf1}
[D]1r4k1/q4bbp/p2p1p2/3P4/4NPQ1/2N4P/PP3RP1/6K1 w - - 7 1
Here I was sure White could take on f6! Apparently not..
35. Kf1 {2,27/10 34 Qe7 Re2} Qe7
{1,98/10 28 Re2} 36. Re2 {2,41/10 31 Qd8} Qd8
{2,15/10 41 Ng3 Qf8 Nf5 h5 Qxg7+ Qxg7 Nxg7} 37. Ng3 {2,49/11 58 Bg6 Nf5} Qf8
{2,41/12 63 Nf5 Kh8} 38. Nf5 {2,79/10 24 Kh8 Nxg7 Qxg7 Qd7 Qf8 Re7 Bg8} Kh8
{2,61/11 37 Nxg7 Qxg7 Qd7 Qf8 Re7 Bg8 Qxd6} 39. Ne4
{2,66/10 35 Rd8 Neg3 Rc8 Nh5 Bxh5 Qxh5 a5} Rd8 {2,45/11 29 Neg3 Rb8 Nh5} 40. b3
{2,71/10 42 Bg6 Re1 Bxf5 Qxf5 Re8} Bg6 {2,29/10 25 Re1 a5 Neg3 Re8 Rxe8 Qxe8
Nxd6}
41. Re1 {2,75/10 28 a5 Rc1 Rc8 Rc4 Rxc4 bxc4 Qb8} a5 {2,62/11 34 g3} 42. g3
{2,91/10 34 Qf7 Nfxd6 Qxd5 Rd1 Qg8 f5} Bf7 {2,50/10 25 Rd1 h5} 43. Nxg7
{2,94/10 38 Qxg7 Qf5 Re8 Rc1 Rxe4 Qxe4} Qxg7 {3,05/10 15 Qf5 Qg6 Qxf6+ Qxf6} 44.
Qf5
{3,10/11 39 Qg6 Qxf6+ Qxf6 Nxf6} Rg8 {2,73/11 30 g4} 45. Kf2
{3,26/11 40 Rd8 Rc1 Qg6 Qxf6+ Qxf6} Qg6 {3,05/11 48 Qxf6+ Qxf6 Nxf6 Rg6 Ne8 Rh6
h4}
46. Qxf6+ {3,46/12 56 Qxf6 Nxf6} Qxf6 {3,60/12 30 Nxf6 Rg6 Ng4 Bxd5 Re8+ Rg8
Rxg8+}
47. Nxf6 {3,25/9 2 Rg6 Ne8 Rh6 h4 Bxd5 Rd1 Be6} Rg6
{3,14/12 21 Ne8 Rh6 h4 Kg8 f5 Kf8 Nc7} 48. Ng4
{3,58/12 37 Bxd5 Re8+ Kg7 Re7+ Bf7 Ra7 a4} Bxd5 {3,34/12 27 Re8+ Kg7 Re7+} 49.
Re8+
{3,65/12 39 Kg7 Re7+ Kh8 f5 Rg5 Ne3 Bc6} Rg8
{3,24/11 28 Nf6 Rxe8 Nxe8 Be6 Nxd6 Bxh3 Kf3} 50. Nf6 {3,86/12 24 Rxe8 Nxe8 h6 g4
Bc6}
Rxe8 {3,66/13 41 Nxe8 Be6 Nxd6 Bd5 Ke3 Kg7 Kd4} 51. Nxe8 {3,50/9 3 h6 g4 Bc6}
Be6
{3,48/12 19 Nxd6 Bd5 Ke3 Kg7 Kd4 Bg2} 52. Nxd6
{4,09/13 44 Bd5 Ke3 Kg7 Kd4 Bg2 Ke5 Bxh3} Bd5 {3,69/12 17 f5 Kg7} 53. Ke3
{4,21/12 23 Kg7 Kd4 Bf3 Nc4 a4 bxa4 Kf6} Kg7 {4,00/12 26 Kd4 Bg2} 54. Kd4
{4,52/12 27 Be6 f5 Bd7 Ke5} Bg2 {4,38/13 40 h4 Kf6 Kc5 Bf3 Kb5 Ke6 Nc4} 55. h4
{4,66/12 33 Kf6 Kc5 Bf1 a3 Ke6 Nc4 Bh3} Kf6 {4,36/12 24 Kc5 Bf1 a3 Bd3} 56. Nc4
{4,85/12 40 a4 b4 Bc6 Kc5 Be4 b5 a3} a4 {4,67/13 37 b4 Ke6 b5 Bh3 Kc5 Kd7 Nd6}
57. b4
{5,14/12 35 Ke6} Ke6 {4,83/12 26 b5 Bf3 b6 h5 Kc5 Bg2 f5+} 58. b5 {5,27/11 19 h5
b6} Bh3
{5,00/11 21 Kc5 Kd7 Nd6 Be6 a3 Bb3 b6} 59. Ne3 {6,47/12 38 Kd6 f5 Bxf5 Nxf5+
Kc7} Kd6
{7,01/12 25 f5 Bxf5 Nxf5+ Kc7 Ng7 Kb6 Kc4} {Score threshold reached} 1-0
(Very) Experimental Betazoid settings:
[Personality = Betazoid 5so] * Rebel 12 further positional tuning
[Pawn Value = 100]
[Knight Value = 101]
[Bishop Value = 100]
[Rook Value = 101]
[Queen Value = 95]
[King Safety = 150]
[Mobility = 120]
[Pawn Structure = 105]
[Passed Pawns = 100]
[Pins = 100]
[Bishop Pair = 100]
[Chess Knowledge = 200]
[Attractiveness = 135]
[Attacking = 130]
[Strength of Play = 100]
[Draw Contempt Factor = 0.00]
[Selective Search = 130]
[Search Technique = NULLMOVE]
[Pruning = MISC_01] * TREE: if (reduce & rd=0) goto tr76
[Pruning ? MISC_25]
[Pruning = MISC_37]
[Pruning ? MISC_40]
[Pruning = MISC_22] * Somewhat more safe Selective Search
[Pruning = MISC_09]
[Pruning = MISC_10]
[Pruning = MISC09_2.00]
[Pruning = MISC10_4.00]
[Pruning = MISC_12]
[Pruning = SEX_HIST_DEPTH_9] * 4
[Pruning = MISC_03] * Move ordering (safe) (speedup 4%)
[Pruning = MISC_46]
[Pruning = M46_HASH_ON]
[Adjusted Selectivity MIDG = 4] * Maximum selective depth (midgame)
[Adjusted Selectivity END0 = 4] * Maximum selective depth (early endgame)
[Adjusted Selectivity END1 = 4] * Maximum selective depth (endgame)
[Adjusted Selectivity END2 = 3] * Maximum selective depth (simple endgame)
[ANTI-GM = OFF]
[EVALUATION = NORMAL]
[Extensions (remaining)= 3]
[Extensions (checks)= 0]
[Extensions (captures)= 0]
Don't try this at home!
I would be happy if I could make some settings for Rebel with a little increased
[Pawn Structure], as that way of playing reminds me a little of my own style. I
am rather bad at tactical positions even though I often play moves like g2-g4 or
h2-h4. But a good pawn structure has a certain beauty of its own.
The pawn structure is like the bones of a position!
Eelco
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.